The result: The change in average sales during the experiment was 7 percent higher at the stores subject to the new policies than at the stores in the control group.
“They were not operating in the stability sweet spot,” said Joan Williams of the University of California Hastings College of the Law, one of the study’s authors. “We basically held up a mirror to capitalism’s self-image of efficiency and showed the misaligned incentives that are disserving both workers and the company.”
Shifting schedules have long been a complaint in the retail industry and other low-wage fields, and the Gap study suggests that they are a lose-lose proposition. But there was one apparent conundrum at the heart of the findings: While the effect on sales was quite large — the 7 percent increase swamps the 1-to-3 percent improvement that many retailers struggle to achieve — the overall consistency of schedules improved only marginally.
Consider a part-time worker with an average of 16 shifts per month. At stores that attempted to make the changes, nearly 10 of those shifts were scheduled for a consistent time of day. At stores that had not done so, about nine of the shifts were consistent.
The explanation for how such modest changes could lead to such unusually large increases in sales may be that managers focused disproportionately on more experienced employees when it came to stabilizing the number of hours worked from week to week. That, in turn, appears to have lowered turnover among experienced workers, who helped their stores perform better.
“It looks like managers did the rational thing — they had to improve stability, and so they picked the more experienced people,” said Saravanan Kesavan, another author, who teaches business at the University of North Carolina. “Now you have more experienced people in the store, and the productivity you’re seeing increased, leading to higher sales.”
Mr. Kesavan suggested that Gap could reap still greater financial benefits by making schedules even more stable and reducing turnover beyond what the experiment achieved. “The attrition rate is still high,” he said. (Susan Lambert of the University of Chicago was the third author.)
The researchers chose the 28 stores in the study, then randomly selected the 19 where the new policies would take effect for just under a year. Although the managers in those stores had previously agreed to take part, the researchers could not be certain that they carried out all the new policies. The managers were simply urged to do so with the company’s blessing, an approach known as “randomized encouragement design.”
Marshall L. Fisher, a professor at the Wharton School at the University of Pennsylvania who is an expert on retailing and was not involved in the project, said the study provided compelling evidence that “if you treat people decently, you get better results.”
Mr. Fisher cautioned, however, that generalizing from the study to the company writ large, to say nothing of the entire industry, could be complicated. For one thing, he said, it was unclear if the improvement in sales would be as large as the researchers found if the company were to put the new policies in place in all of its stores at once. It is possible that the stores that adopted the new policies became attractive to shoppers, and may have siphoned traffic and sales from other Gap stores.
In a statement, Gap was noncommittal on whether it would introduce most of the changes across all of its stores.
Looming over the study was a gnawing question: Given the effort that went into making workers’ schedules more consistent from day to day and week to week, why wasn’t there more improvement in the actual scheduling?
One answer, the researchers concluded, is that the biggest cause of disruption in work schedules at Gap stores wasn’t the fluctuation in traffic from customers, as was long assumed to be the case in retailing. Overly narrow decision-making at the corporate headquarters was a major factor.
In 2015, as the study was beginning, Gap announced it would require schedules to be posted at least 10 days in advance at all its stores in the United States. But schedules can nonetheless change quickly. According to interviews the researchers conducted with workers and managers, sales promotions mandated by corporate headquarters often required a large increase in worker-hours on short notice. So did preparing for store visits by executives.
“It’s just been a roller coaster with promo changes,” one manager told the researchers. “This week alone the window changed three times.”
Another manager reported having “probably extended two to three shifts every day in the run up to the visit” by an executive.
The researchers also identified inaccurate information about the size and timing of shipments as a key obstacle to predictable scheduling, since processing new inventory is labor-intensive.
Mr. Fisher said those findings could have an optimistic interpretation, because they suggested that it was largely within the company’s power to smooth out the peaks and troughs in its workers’ schedules. Some companies, like Costco, are able to provide more regular schedules in part by stocking a smaller range of products or scaling back or eliminating promotions.
In a statement, Gap brand’s top human resources official, C. David Ard, said, “We are taking a closer look at certain decisions made at headquarters and actively examining changes that could further stabilize the store scheduling process.”
In the meantime, the fact that an inconveniently timed shipment could derail a store’s schedule may point to a broader issue: Stores like Gap frequently leave little margin for error when it comes to staffing because managers are given targets for labor hours and their ability to meet them figures prominently in their evaluations.
“A lot of the instability comes from managers having tight labor budgets,” said Carrie Gleason, the director of the Fair Workweek Initiative at the Center for Popular Democracy, an advocacy group. “They’re understaffing to basically meet their numbers from one week to the next. It’s management by the practice of triage.”
As part of its scheduling changes in 2015, Gap officially eliminated tentative “on-call” shifts that could be canceled at the last minute. But Ms. Stephenson, the Gap worker in North Carolina, said it was not uncommon for her managers to cancel a shift on short notice when they believe they can get by without an extra worker.
“They try to do the best with what they’ve got, and they’re very apologetic,” Ms. Stephenson said. “They say, ‘Corporate says we don’t have the hours.’”
Mr. Ard said: “While we can’t speak to this particular employee’s experience without further investigation, we will be looking into this issue as we treat such concerns seriously.”
The most important contribution of the Gap study may be to chip away at a rigid approach to labor costs that may not always reflect a rational calculation.
“It is massively useful,” said Ms. Gleason of the Fair Workweek Initiative. “It helps make the business case for why stable schedules work.”